[sword-devel] Re: [bt-devel] fast search

David Twyerould sword-devel@crosswire.org
Wed, 14 Feb 2001 18:35:54 +1100


David/Trevor, I'm not sure if this helps, but on the subject of inverted
indexes, my understanding is that the Online Bible format is in fact an
inverted index with punctuation. From memory I think there are three parts
to it: the words themselves(word list), the word index and the verse list
(ie: a sequential list of word numbers).  The verse text is recreated on
the fly as required. The index is compressed as is the word list.

The advantage is that you have minimal disk space plus high speed
searching. The time taken to reconstruct the verse text is negligable on
all but the very slowest machines and even then its not usually a problem
(I'm talking Intel 286 machines with Windows 3.1).

David Twyerould